ClawHub
Back to skill

Security audit

可逆关键词脱敏

Security checks for vulnerabilities and agentic risk

Overview

This skill is a local document masking tool with disclosed file processing and no evidence of network exfiltration or hidden persistence.

Install only if you are comfortable with a local CLI reading and writing documents in its working directory. Keep mapping files, password files, and RKM_KEY private and out of Git. For documents from untrusted sources, prefer .txt/.md/.docx where possible or run legacy .doc conversion in a sandboxed/offline environment because Office and LibreOffice parsers have their own attack surface.

SkillSpector

Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

subprocess module call

Medium
Category
Dangerous Code Execution
Content
env["RKM_WORD_OUTPUT"] = str(output_path)
    env["RKM_WORD_FORMAT"] = str(file_format)
    creationflags = getattr(subprocess, "CREATE_NO_WINDOW", 0)
    result = subprocess.run(
        [powershell, "-NoProfile", "-NonInteractive", "-ExecutionPolicy", "Bypass", "-Command", script],
        capture_output=True,
        text=True,
Confidence
91% confidence
Finding
This code launches PowerShell to automate Word COM for .doc/.docx conversion. Although it uses an argument list rather than shell string concatenation, invoking an external interpreter with ExecutionPolicy Bypass and processing untrusted document paths expands the attack surface and can trigger risky behavior in local Office automation or document parsing components.

subprocess module call

Medium
Category
Dangerous Code Execution
Content
target_ext = output_path.suffix.lower().lstrip(".")
    with tempfile.TemporaryDirectory(prefix="rkm-lo-") as tmp:
        tmp_dir = Path(tmp)
        result = subprocess.run(
            [executable, "--headless", "--convert-to", target_ext, "--outdir", str(tmp_dir), str(input_path)],
            capture_output=True,
            text=True,
Confidence
88% confidence
Finding
This code executes LibreOffice/soffice on attacker-controlled document input. Even without shell injection, handing untrusted files to large external parsers is a real security boundary crossing and can expose the host to parser exploits, unexpected network/file behaviors, or malicious embedded document content handled by the converter.

Context-Inappropriate Capability

Medium
Confidence
90% confidence
Finding
The skill's stated purpose is local masking/restoration, but this section adds external document conversion via Word COM and LibreOffice. That extra capability materially broadens the trust boundary and exposes the host to third-party parsers and automation engines that may process hostile .doc content.

Static analysis

No suspicious patterns detected.